How AI Optimizes Water Supply: Pumps, Leaks, and Quality Control

We develop AI systems for water supply that reduce costs and improve reliability. Water supply is one of the most energy-intensive municipal systems—typical non-revenue water (NRW) in Russian cities ranges from 20–40%, and pumps consume up to 80% of the entire water utility's electricity. Traditional control methods—manual operation or PID controllers—do not account for tariff schedules or demand forecasts. We use AI to cut these costs: we optimize pump operation considering tariffs, detect leaks without excavation, and monitor water quality in real time. Our engineers have experience with over 10 projects in municipal utilities, guaranteeing a 15–25% reduction in electricity costs, often translating to over $50,000 in annual savings for a typical utility.

How AI Reduces Electricity Costs

Pumps are the main energy consumers. Optimizing their schedule with respect to tariff rates is a classic linear programming problem. We solve it over a 24-hour horizon: shifting pump operation to nighttime hours when tariffs are lowest, while maintaining pressure and reservoir levels.

import numpy as np
from scipy.optimize import minimize
import pandas as pd

class PumpScheduleOptimizer:
    """Pump schedule optimization considering tariff schedule"""

    def __init__(self, n_pumps, reservoir_capacity_m3):
        self.n_pumps = n_pumps
        self.V_max = reservoir_capacity_m3
        self.V_min = reservoir_capacity_m3 * 0.2  # min 20% capacity

    def optimize_24h(self, demand_forecast, tariff_schedule, pump_specs, V_init):
        """
        demand_forecast: hourly water demand [m³/h]
        tariff_schedule: hourly electricity tariff [rub/kWh]
        pump_specs: [{flow_m3h, power_kw, min_run_time_h}]
        V_init: initial reservoir level [m³]
        """
        T = 24  # 24-hour horizon

        from pulp import LpProblem, LpMinimize, LpVariable, lpSum, LpBinary

        prob = LpProblem("pump_schedule", LpMinimize)

        # Binary variables: pump i on at hour t
        pump_on = [[LpVariable(f"pump_{i}_{t}", cat='Binary')
                    for t in range(T)] for i in range(self.n_pumps)]

        # Reservoir level variable
        V = [LpVariable(f"V_{t}", lowBound=self.V_min, upBound=self.V_max)
             for t in range(T+1)]
        V[0].setInitialValue(V_init)

        # Objective: minimize electricity cost
        prob += lpSum(
            pump_specs[i]['power_kw'] * pump_on[i][t] * tariff_schedule[t]
            for i in range(self.n_pumps) for t in range(T)
        )

        # Reservoir balance
        for t in range(T):
            inflow = lpSum(pump_specs[i]['flow_m3h'] * pump_on[i][t]
                          for i in range(self.n_pumps))
            prob += V[t+1] == V[t] + inflow - demand_forecast[t]

        prob.solve()
        return [[pump_on[i][t].value() for t in range(T)] for i in range(self.n_pumps)]

Optimization effect: By shifting pump operation to nighttime hours (cheap tariff), savings amount to 15–25% of electricity costs for the same water volume. Unlike replacing pumps with more efficient ones (which requires capital investment), AI optimization uses existing equipment and pays back in 3–6 months.

How to Detect Leaks Without Excavation

We combine three methods:

• Balance analysis: compare supplied and consumed water volumes to identify zones with abnormally high NRW.
• Pressure transient analysis: LSTM on pressure time series detects bursts and slow leaks.
• Correlation method: two acoustic sensors with time difference of arrival determine leak distance with accuracy of 50–100 m.

Comparison with traditional methods: AI approaches detect leaks 3–5 times faster and require 2 times fewer crew dispatches. Traditional methods like correlators and thermal imagers require prior assumptions about leak location and provide accuracy no better than 200 m.

Water Quality Control Using ML

IoT sensors measure chlorine, turbidity, pH, and temperature in real time. We build a hydraulic model using EPANET and predict chlorine concentration at any point in the network. This allows maintaining SanPiN standards (>0.05 mg/L) with minimal dosing, saving up to 30% reagent.

Contamination detection: anomalous parameter combinations (e.g., turbidity + pH) trigger an ML classifier that indicates the likely source and spread zone.

Sewer Network Management

For sewer networks, we predict wastewater inflow during storms using LSTM on rain gauge data and collector levels. A 1–2 hour forecast enables proactive gate adjustments to avoid overflow. Optimization of sewer pumping stations smooths peaks, preventing overload at treatment plants.

Why AI Optimization Is Better Than Pump Replacement

Replacing pumps with more efficient ones requires capital expenditure and prolonged downtime. AI optimization uses existing equipment, adapting its modes to current conditions. Comparison:

Development Process and Timelines

A typical project takes 4–7 months and includes the following stages:

1. Data audit (2–3 weeks): collection and validation of historical data.
2. Modeling (4–6 weeks): hydraulic network model in EPANET, ML model training.
3. Integration (3–5 weeks): connection to SCADA, IoT platform, dashboard setup.
4. Testing (2–3 weeks): A/B test on real data, validation of savings.
5. Deployment and training (2–4 weeks): container installation, staff training.

What Is Included in the Work

• Analytical report with current losses and savings potential.
• Hydraulic network model (EPANET).
• ML models (pump optimization, leak detection, water quality).
• Integration with existing systems (SCADA, IoT platform).
• Dashboards for dispatchers and reports for management.
• Staff training and documentation.
• 6-month warranty support.

Why Choose Us

We have developed AI systems for 5 water utilities (over 10 projects). We guarantee a 15–25% reduction in electricity costs and a 10–30% reduction in water loss. We use only proven stacks: PyTorch, LSTM, LP-solvers. We provide post-support and adaptation to your specific network.

Want to assess the potential of your system? Contact us for a 1-week audit and a proposal. Order a pilot project: results are visible within 2 months of start.

SanPiN 2.1.3684-21 — drinking water quality standards

Industry AI Solutions: Healthcare, Finance, Retail, Manufacturing

We encounter the same pain points: a general text model doesn’t distinguish medical nomenclature, and a standard object detector confuses “weld seam scratch” with “casing scratch.” Each time these are different defects with different consequences. To avoid this, we build industry-specific solutions on top of general methods, but with deep domain knowledge — from regulatory requirements to data specifics. Over 5 years, we have completed 80+ projects in fintech, healthcare, retail, and manufacturing, and none were without adaptation to a specific business case.

Healthcare: Regulatory Maze and Data Governance

Medical AI differs not in technical algorithms but in a compliance-first approach. Depending on the country of application, the model may be a Class II or III medical device requiring clinical trials (FDA, CE MDR, GOST R). We ensure compliance with these standards at the architecture stage — fixing them post-factum is 10× more expensive.

Medical imaging. Detection on X‑rays, CT, MRI is a mature area. Models on ResNet, EfficientNet, SegFormer achieve AUC 0.94–0.97 on standard tasks (pneumonia on CXR, polyps on colonoscopy). Key issue is generalization: a model trained on data from one scanner manufacturer degrades on another due to differences in preprocessing and artifacts. Solution: domain adaptation via MONAI (Medical Open Network for AI) from NVIDIA, which includes DICOM loading, 3D augmentation, and confidence calibration. TotalSegmentator — for automatic segmentation of 117 structures on CT, production‑ready, Apache 2.0 license.

Clinical NLP. Extracting structured information from clinical records: diagnoses (ICD‑10/11), prescriptions, dates, indicators. medspaCy, scispaCy, MedCAT — specialized NLP libraries with ontologies (SNOMED‑CT, UMLS). Fine‑tuning BioBERT or ClinicalBERT on our data yields F1 0.85–0.92 on NER tasks versus F1 0.65–0.72 for general BERT. We verified this on a project with a regional oncology center — cancer stage extraction accuracy increased by 23%.

Clinical decision support. LLM assistants for clinical decision support are a regulatory gray area. We use an RAG system on top of clinical guidelines (UpToDate, local protocols) with explicit citation for each statement. The model does not diagnose but helps find relevant protocols. Stack: LlamaIndex + pgvector + pubmedbert-base-embeddings + Llama Guard for safety. Data in DICOM/HL7 FHIR, on‑premise deployment mandatory.

Finance: How to Ensure Interpretability of a Scoring Model under Basel IV?

The financial sector is one of the most mature in applying ML, but regulation is maximal. Every model affecting credit decisions falls under Basel IV, EU AI Act, GDPR Article 22. We deliver AI solutions for fintech that satisfy these requirements — in a project for a top‑10 bank we deployed a scoring model where each record required SHAP explanations.

Credit scoring. Gradient boosting (LightGBM, XGBoost) dominates. Neural networks yield +0.5–2% AUC but lose interpretability. Standard: LightGBM + SHAP to explain each decision. Fairness checking is mandatory: Fairlearn or aif360 for auditing disparate impact on protected attributes (age, gender). The default class is 1–5% — with an imbalance of 1:30, a model with 97% accuracy may have recall 0.2. Solution: focal loss, class_weight='balanced', SMOTE + careful validation. In one fintech scoring project, the model reduced credit losses by $2.1 million annually.

Algorithmic trading and risk management. LSTM and Transformer for price forecasting are popular but unstable in production due to non‑stationarity of financial series. A more robust approach: ML for signal generation (classification: up/down over horizon N) with traditional portfolio optimization on top. Backtesting via Zipline‑Reloaded, vectorbt, QuantLib. Proper backtesting is critical — look‑ahead bias kills results. We guarantee a clean experiment: all data at signal time is available in real time.

AML (Anti‑Money Laundering). Graph Neural Networks for analyzing transaction networks is an actively developing area. PyG, DGL for GNN. Task: detect suspicious patterns in transaction graphs (layering, structuring). Recall is more critical than precision — better 10 false alarms than miss one money laundering. In a project for a large payment service, we increased recall by 18% without increasing false positive rate.

Retail and e‑commerce: Recommendation Systems and Demand Forecasting

Recommendation systems. Current architectural standard: two‑tower model for retrieval + ranking with cross‑features. TensorFlow Recommenders or Merlin from NVIDIA for GPU‑accelerated feature processing. For small catalogs (<100k items), LightFM is sufficient. A common mistake is training on implicit feedback without accounting for position bias. Solution: IPW (Inverse Propensity Weighting) or randomized logging on a portion of traffic. Development time for a basic recommendation system is 4–8 weeks, including A/B test.

Demand forecasting and inventory optimization. Hierarchical forecasting: SKU → category → store → region. HierarchicalForecast from Nixtla automatically reconciles forecasts across levels. TFT or N‑HiTS for base forecast, gradient boosting for adjustment on exogenous factors (promotions, weather, events). One retail project led to a 15% reduction in stock‑outs due to precise promotion calibration.

Visual search and size compatibility. CLIP embeddings for image search — deploy in 2–3 weeks: clip‑ViT‑B‑32 or clip‑ViT‑L‑14, Faiss or Qdrant index, REST API. For size recommendation — specific models on return data and reviews with fit indication.

Manufacturing: Quality Inspection and Predictive Maintenance

Quality control and defect detection. CV models for product inspection are one of the most mature industry tasks. YOLOv10 for defect detection, SegFormer for segmentation. Specifics: class imbalance (defects are rare), high recall requirement (missing a defect is worse than false alarm). Typical dataset: 500–2000 defect images + 500–1000 normal. Few‑shot learning via DINO or SAM 2 works with 50–100 annotated examples. We gained experience on an electronics production line — recall 0.95 at FPR 0.03. A predictive maintenance deployment saved a manufacturing client $500,000 per year in unplanned downtime.

Predictive maintenance. Vibration sensors, current sensors, thermocouples → feature extraction → anomaly or mode classification. Models: LSTM‑AE for unsupervised, LightGBM for supervised (if failure history is available). Integration with SCADA/OPC‑UA via opcua-asyncio or MQTT. Key metric: False Negative Rate — a missed pre‑failure is more costly than a false alarm. Threshold tuned to business cost of each error type. Timeline: 3 to 6 months to production.

Digital twin and simulation. Surrogate models — ML models replacing expensive physical simulation. If a CFD simulation takes 6 hours and a surrogate (trained on 10,000 simulations) takes 0.01 seconds, that's 2,000,000× speedup for optimization. SALib for sensitivity analysis, botorch for Bayesian optimization on top of surrogate.

General Principles of Industry AI

Regardless of industry, there are patterns that work everywhere. Data matters more than architecture. In healthcare, 1000 quality labeled images are better than 100,000 poor ones. In manufacturing, 200 real defect examples are more valuable than 10,000 synthetic ones. Compliance‑first design — regulatory requirements are easier to embed into architecture from the start than to add later. Logging, explainability, versioning from day one. Domain expert on the team — an ML engineer without domain knowledge does slowly and error‑prone what an ML engineer plus a doctor/financier/technologist does quickly and correctly.

We guarantee certification to customer requirements (ISO 13485, SOC 2, GDPR) and provide full model documentation (model card, datasheet, compliance report). Our experience: 10,000+ engineering hours and 80+ projects.

Work Process for an Industry AI Solution

1. Domain immersion (2–3 days) — interviews with experts, studying regulatory requirements, auditing available data.
2. MVP design (1–2 weeks) — stack and architecture selection, feasibility assessment.
3. Development and validation (from 4 weeks to 6 months depending on industry) — model training, testing, compliance.
4. Integration and deployment (1–4 weeks) — on‑premise / cloud / edge, documentation, staff training.
5. Support and monitoring — model drift, retraining, SLA.

Estimated timelines:

Cost is calculated individually for each project. Get a consultation — we will evaluate your dataset, regulatory map, and business goals.

Why Choose Our Industry AI Solutions?

• 80+ completed projects in fintech, healthcare, retail, and manufacturing.
• 5 years on the market — proven experience with compliance and deployment.
• Quality guarantee: we ensure target metrics (AUC, recall, latency p99) and provide full documentation.
• Licensed technologies: PyTorch, MONAI, LightGBM, Qdrant — we use open‑source with commercially safe licenses.
• Flexibility: we work as a contractor or as an extension of your team.

Contact us for a free data audit and consultation. Request a proposal with a detailed work plan. We will discuss your task and prepare a commercial proposal.